ARTIFICIAL INTELLIGENCE POSSIBILITIES IN ASSESSING STUDENTS’ INDEPENDENT WORK IN HIGHER EDUCATION

Authors

  • Sultanbek Nazerbaevich Kalbaev

    Nukus State Pedagogical Institute named after Ajiniyaz image/svg+xml

Keywords: independent work, assessment, artificial intelligence, mathematical model, formative feedback, reliability, validity, mutual information, equilibrium, rubric

Abstract

The article analyses the possibilities of artificial intelligence (AI) in assessing students’ independent work from a mathematical standpoint: notation is introduced, seven postulates are stated formally, and a sequence of theorems is derived with proofs. It is shown that the question “is AI useful” is ill-posed; the correct question is where the binding constraint of learning through independent work lies and whether the cost structure of AI relaxes it without making its own errors expensive. From the postulates it is proved that the binding constraint is the feedback bottleneck, that the natural domain of AI is the formative function, that reliability and validity are orthogonal, that the quality of AI assessment is bounded above by rubric precision, and that the unit of assessment necessarily shifts from product to process.

References

1. Tupikova S.E., Bykova N.O. Chet tillari fakultetlari talabalarining mustaqil ishini tashkil etishda sunʼiy intellekt imkoniyatlari. // Toʻplam materiallari. Saratov: N.G. Chernishevskiy nomidagi Saratov milliy tadqiqot davlat universiteti, 2023. 214-221-b.

2. Oʻquv jarayonida sunʼiy intellektdan foydalanish imkoniyatlari (oliy taʼlim misolida) //CyberLeninka [Elektron resurs]. URL: https://cyberleninka.ru/article/n/o-quv-jarayonida-sun-iy-intellektdan-foydalanish-imkoniyatlari-oliy-ta-lim-misolida (murojaat sanasi: 01.07.2026).

3. McCarthy J. et al. A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence. 1956 – har ikki manba tayangan sunʼiy intellekt taʼrifining birlamchi manbai.